jz88rh56ms-eng

Kraken is a high-performance, stateful **Agent Operating System** built in Rust and powered by WASM.

10
0
89% credibility
Found May 17, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

Kraken is an experimental platform for running portable software programs (called agents) in isolated sandboxes. These agents can store information, send signed messages to each other across a peer-to-peer network, and even call AI assistants. The key feature is that agents can be moved between computers while keeping their identity and state intact through encrypted snapshots. It's built in Rust for reliability and targets WebAssembly for portability and safety.

How It Works

1
🔬 You hear about an interesting experiment

A developer shares a project called Kraken that can run AI programs as portable, migratable agents on any computer in a network.

2
🚀 You run the demo and watch two programs discover each other

With one command, two invisible programs running on your machine find each other automatically, exchange information about what they can do, and introduce themselves.

3
💬 One program sends a signed message to the other

A tiny program (called a 'pinger') signs a message with its own digital identity and sends it across the network to another program (called an 'echo').

4
You explore what you can build
🎮
Build your own agent program

Write a program in Rust that gets loaded into the sandbox with specific permissions to store data, send messages, or talk to an AI assistant

🔄
Move an agent to another computer

Take a running program, package its entire state (memory and identity intact), and send it to another node in the network

5
🔐 Your agent arrives safely with its secrets preserved

The receiving computer verifies the package is genuine (using the agent's digital signature) and restores the program with all its state exactly as it was.

Everything works end-to-end

Your agent continues running on the new computer with its counter showing the same number, ready to receive new messages and work as before.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is Kraken?

Kraken is a stateful agent runtime built in Rust that lets you run AI agents as WASM modules. Think of it as an operating system for AI agents -- each agent runs in a sandboxed environment with its own cryptographic identity and persistent state. The key innovation is that agents can migrate between nodes while preserving their state and identity, using a P2P protocol built on libp2p. It supports OpenAI, Anthropic, Gemini, and OpenRouter-compatible endpoints for LLM calls.

Why is it gaining traction?

The combination of WASM sandboxing with cryptographic identity and state migration is novel. Agents maintain their state and cryptographic keypair across node migrations, which is valuable for distributed systems. The capability-based security model is more granular than typical agent frameworks, and the signed messaging with replay protection provides integrity guarantees. The demo runs two nodes that auto-discover each other, exchange signed messages, and migrate agents with state preserved -- a compelling proof of concept.

Who should use this?

This is for developers building distributed AI systems who need agents that can move between nodes without losing context. Researchers exploring sovereign agent runtimes and multi-agent orchestration would find this useful. Teams requiring cryptographic guarantees around agent communication and identity will benefit most.

Verdict

Kraken shows architectural promise with 57 passing tests and a well-documented design, but the 0.9% credibility score and 10 stars reflect an early-stage project. The v0.1 alpha status means the API is unstable and documentation is limited. Worth exploring for experimentation, but production use requires caution.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.